This model is shockingly small for how capable it is. its a little bit bigger than deepseekV4 flash but around as capable if not more on some benchmarks than V4 pro, i wouldnt be surprised if this becomes a popular local model.

▲

andai 2 hours ago | parent | next [-]

I've been wondering about that. GLM-5.2 is also half the size of DeepSeek V4 Pro. (But costs roughly twice as much.)

I looked into DeepSeek's architecture a little bit and the main focus was how can we save as much money as possible. They did a lot of cost cutting with the attention mechanisms. This allowed them to offer an insanely cheap price even on massive contexts, but seems to have come at the cost of performance?

At least, that's my guess, when I see smaller models costing more and outperforming, I think, "they must have denser attention?"

	▲	wgd an hour ago \| parent [-]
		The current Deepseek V4 Pro is still just their initial preview AFAIK, with the "real" model release rumored to come later this month. GLM-5.2 might be outperforming simply because it's had more post-training on top of the GLM-5 base.

▲

nunodonato 2 hours ago | parent | prev | next [-]

hardly, its still quite big unless by "local" you mean people that spend many thousands on rigs :)

	▲	nshotton 2 hours ago \| parent [-]
		Yeah i shouldve been more clear, a model of this size could run on 2 dgx sparks so out of the range of a lot of the typical consumer sure, but I think there is definitely a market for that size

▲

IshKebab 2 hours ago | parent | prev [-]

> Hy3 has 295B parameters in total. To serve it on 8 GPUs, we recommend using H20-3e or other GPUs with larger memory capacity.

I would.